INDEX

Explanations

politically-themed phrases related to legislation and government officials

oai_token-act-pair · gpt-3.5-turbo

Triggered by @bot

mentions of politicians and their associated party affiliation

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-att-kk

Configuration

ckkissane/attn-saes-gpt2-small-all-layers/gpt2-small_L3_Hcat_z_lr1.20e-03_l19.00e-01_ds24576_bs4096_dc1.00e-06_rsanthropic_rie25000_nr4_v9.pt

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

float32

Hook Name

blocks.3.attn.hook_z

Hook Layer

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Activation Function

relu

Embeds

IFrame

Link

Head Attr Weights

0:0.02

1:0.38

2:0.06

3:0.04

4:0.02

5:0.03

6:0.07

7:0.08

8:0.06

9:0.05

10:0.06

11:0.09

Negative Logits

DragonMagazine

-1.96

Reviewer

-1.61

 Solitaire

-1.54

ijn

-1.52

��

-1.51

PDATE

-1.45

 converter

-1.42

alys

-1.38

 Inquisition

-1.33

quished

-1.32

POSITIVE LOGITS

)]

1.68

pel

1.53

ulla

1.43

 Democr

1.41

anan

1.41

uay

1.39

Washington

1.39

oyer

1.37

asse

1.37

Newsletter

1.34

Activations Density 0.035%

politically-themed phrases related to legislation and government officials

mentions of politicians and their associated party affiliation

No Known Activations